lable at ScienceDirect

Applied Geography 67 (2016) 77e93

Contents lists avai

Applied Geography

journal homepage: www.elsevier .com/locate/apgeog

Stakeholder analysis for marine conservation planning using publicparticipation GIS

Greg Brown a, *, Jennifer Strickland-Munro b, Halina Kobryn b, Susan A. Moore b

a The University of Queensland, Australiab Murdoch University, Australia

a r t i c l e i n f o

Article history:Received 30 October 2015Received in revised form7 December 2015Accepted 14 December 2015Available online xxx

Keywords:Stakeholder analysisConservationPPGISMarine protected areas

* Corresponding author.E-mail addresses: greg.brown@uq.edu.au (G. B

murdoch.edu.au (J. Strickland-Munro), H.Kobryn@muMoore@murdoch.edu.au (S.A. Moore).

http://dx.doi.org/10.1016/j.apgeog.2015.12.0040143-6228/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

Stakeholders are presumed to represent different interests for marine and coastal areas with the po-tential to influence marine protected area planning and management. We implemented a publicparticipation GIS (PPGIS) system in the remote Kimberley region of Australia to identify the spatial valuesand preferences for marine and coastal areas. We assessed similarities and differences in PPGIS partic-ipants (N ¼ 578) using three operational definitions for “stakeholder” based on: (1) self-identified group,(2) self-identified future interests in the region, and (3) participant value orientation that reflects apreferred trade-off between environmental and economic outcomes. We found moderate levels of as-sociation between alternative stakeholder classifications that were logically related to general and place-specific participatory mapping behavior in the study region. We then analyzed how stakeholder clas-sifications influence specific management preferences for proposed marine protected areas (MPAs) in thestudy region. Conservation-related values and preferences dominated the mapped results in all proposedmarine reserves, the likely result of volunteer sampling bias by conservation stakeholder interestsparticipating in the study. However, we suggest these results may also reflect the highly politicizedprocess of marine conservation planning in the Kimberley where conservation efforts have recentlyemerged and galvanized to oppose a major offshore gas development and associated land-based infra-structure. Consistent with other participatory mapping studies, our results indicate that the chosenoperational definition for stakeholder group such as group identity versus interests can influenceparticipatory mapping outcomes, with implications for MPA designation and management. Futureresearch is needed to better understand the strengths and limitations of participatory mapping that isframed in stakeholder perspectives, especially when sampling relies heavily on volunteer recruitmentand participation methods that appear predisposed to participatory bias. In parallel, practical efforts toensure that social research efforts such as this are included in MPA planning must remain of the highestpriority for scientists and managers alike.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Marine protected areas (MPAs) are designated to enhance con-servation of marine resources and provide an important tool tocounter the rapid degradation of the world's oceans (Lubchenco,Palumbi, Gaines, & Andelman, 2003). Despite significant growthin recent years, the establishment of MPAs, as a percent of totalmarine area, lags terrestrial protected areas. In 2014, MPAs covered

rown), J.Strickland-Munro@rdoch.edu.au (H. Kobryn), S.

3.4% of the global ocean area, 8.4% of the area under nationaljurisdiction (0e200 nautical miles), and 10.9% of all coastal waters,but only 0.25% of marine areas beyond national jurisdiction (Juffe-Bignoli et al., 2014). In contrast, 15.4% of the world's terrestrialareas, including inland waters, have protected area status (Juffe-Bignoli et al., 2014).

Stakeholders play a critical role in the establishment and man-agement of MPAs which are often political and contentious asillustrated by events in Australia. In 2012, a Labor governmentannounced an additional 2.3 million square kilometers would beadded to the current Commonwealth marine reserve system,bringing the system total to over 3.1 million square kilometers.Marine reserve plans were approved for implementation in 2014,

G. Brown et al. / Applied Geography 67 (2016) 77e9378

but with an electoral change to a Liberal government, the planswere suspended and the government commissioned a review ofthe system. The government stated the review fulfilled an electioncommitment to ensure that “management arrangements for thereserves reflect genuine and thorough consultation with stake-holders and are informed by the best available science”(Department of Environment, 2015).1 Commercial fishing stake-holders were presumed to have played an important role in thegovernment decision to suspend the reserve plans pending review.

There are multiple definitions for stakeholders, but one that fitsthe purpose of this study defines stakeholders as “any group ofpeople, organized or unorganized, who share a common interest orstake in a particular issue or system … who can be at any level orposition in society, from global, national and regional concernsdown to the level of household or intra-household, and be groupsof any size or aggregation” (Grimble & Wellard, 1997, p. 176).Stakeholders can also include the nebulous categories of ‘futuregenerations’, the ‘national interest’ and ‘wider society’ (Grimble &Wellard, 1997), with these categories often evoked as justificationfor the establishment of MPAs. A key distinction between stake-holders is those who affect decisions and those who are affected bydecisions. This distinction has significant implications for stake-holder analysis methods that can identify stakeholder groups priorto the initiation of a planning process, or alternatively provide forthe emergence of stakeholder groups through an inductive analysisof expressed preferences (Brown, de Bie, & Weber, 2015).

There is widespread agreement on the importance of incorpo-rating stakeholders in meaningful participation for effectivemarineconservation planning and management (Charles & Wilson, 2009;Lundquist & Granek, 2005; Pomeroy & Douvere, 2008; Pollnac,Crawford, & Gorospe, 2001, 2010; Voyer, Gladstone, & Goodall,2012), in all phases of marine conservation ranging from marineprotected area design to implementation and management.Stakeholders can assist in the identification of marine spatial planpriorities and objectives, the selection of options, plan imple-mentation and enforcement, and evaluation of outcomes (Pomeroy& Douvere, 2008). MPAs are unlikely to meet their biological orsocial goals unless the human dimensions or people-oriented fac-tors are integrated into the MPA design and evaluation process(Charles & Wilson, 2009; Christie et al., 2003; Gruby, Gray,Campbell, & Acton, 2015; Pollnac et al., 2010). Indeed, some arguethat MPA failure may be attributable to consultative failures in theearly stages when an MPA is conceived, communicated, and dis-cussed among stakeholders (Chuenpagdee et al., 2013). MPA de-signs that include both biodiversity conservation goals andmultiple socioeconomic stakeholder interests are more likely toprotect marine ecosystems (Christie, 2004; Klein et al., 2008), whileMPA management strategies that find the “middle-ground” be-tween government-led and community-based approaches may bemost effective (Jones, 2002).

The purpose of stakeholder analysis is to inform the develop-ment and consideration of alternatives in the early stages of aproject or proposal, or if a project or plan has been implemented, toeffectively manage stakeholders and conflicts over the duration ofthe plan. Stakeholder analysis is particularly relevant for environ-mental issues such as marine conservation because potential im-pacts tend to cross-cut biophysical and social systems, involvemultiple uses and user groups, contain externalities and trade-offs,and affect future availability or productivity of resources (Grimble& Chan, 1995; Grimble & Wellard, 1997). In the application ofstakeholder analysis to marine conservation, stakeholder analysisappears especially important in the early stages of design and

1 https://www.environment.gov.au/marinereservesreview/about.

zoning of MPAs, but stakeholders can also be used to verify evi-dence collected in support of a marine spatial planning process(Shucksmith, Gray, Kelly, & Tweddle, 2014).

The need to identify and understand stakeholders is part ofbroader and increasing urgent calls to include social science in MPAplanning and management. Gruby et al. (2015) advocate forresearch scoping the diverse values of MPAs, while Voyer et al.(2012) focus on social assessment, encouraging researchers tomove beyond public participation. This paper makes an importantcontribution in progressing social research, with a strong spatialfocus, while also extending our understanding of social assess-ments. This contribution involves understanding stakeholders andhow their operational identity affects analysis of planning andmanagement alternatives. Voyer et al. (2012) note the need tomovebeyond a generic perspective on public participation; this paperprogresses our understanding by interrogating who is the “public”and provides methods for doing so.

1.1. Stakeholder analysis methods and participatory mapping

There are a range of methods for identifying and analyzingstakeholder perspectives for environmental planning and man-agement, including marine conservation. For example, Reed et al.(2009) describe three steps in stakeholder analysis: identifyingstakeholders, differentiating between and categorizing stake-holders, and investigating relationships between stakeholders.Grimble and Chan (1995) describe the following steps: identify thepurpose of analysis (goals); develop an understanding of the sys-tem, decision makers, and drivers of decisions; identify principalstakeholders; investigate stakeholder interests, characteristics andcircumstances; and identify patterns and contexts of interactionbetween stakeholders. Stakeholder analysis, as traditionally prac-ticed, identifies key individuals and groups through expert-drivenprocesses that do not usually include broad-based social surveys.For example, the Marine Life Protection Act initiative in Californiathat established a system of marine reserves used a regionalstakeholder group process where stakeholders were identified,appointed, and worked in small, staff-supported groups to developmultiple MPA proposals over the course of about one year (Foxet al., 2013).

The emergence of participatory mapping methods usinggeographic information is a relatively recent addition to thestakeholder analysis toolbox. Public participation geographic in-formation systems (PPGIS), participatory GIS (PGIS), and vol-unteered geographic information (VGI) describe methods thatcommonly engage lay people (non-experts) to generate spatialinformation for a wide range of urban, regional, and environmentalplanning applications (see Brown & Kytt€a, 2014; Brown, 2012,2005). Participatory mapping for environmental applicationsoften identifies place-based values (Brown & Reed, 2000) andplace-based preferences (Brown, 2006). Mapped place-basedvalues and preferences, when combined with participant charac-teristics, provide an alternative approach to common stakeholderidentificationmethods. Most PPGIS/PGIS/VGI processes that informenvironmental planning involve stakeholders given the broaddefinition of stakeholder that includes those affected by planningdecisions. Schlossberg and Shuford (2005) describe how the term“public” in PPGIS can refer to decision makers, implementers,affected individuals, interested observers, or the general publicdinother words, stakeholders.

With participatory mapping, the focus of stakeholder analysisexpands from individuals and groups perceived to havemore directinfluence/power over marine planning decisions to those that arepotentially affected by decisions. These individuals can be termed“latent” stakeholders (Mitchell, Agle, & Wood, 1997) that possess

G. Brown et al. / Applied Geography 67 (2016) 77e93 79

legitimacy for involvement, but not necessarily the power or ur-gency to engage with the process. Participatory mapping can alsoexplicitly sample for “definitive” stakeholders (Mitchell et al., 1997),that is, those possessing power, legitimacy, and urgency forinvolvement. In one of the few examples of participatory mappingfor marine spatial planning, Ruiz-Frau, Edwards-Jones, and Kaiser(2011) mapped stakeholders' values for marine ecosystems andassessed their preferences for the location and type of marineprotected areas (MPAs) around the coast of Wales (UK). Individualsbelonging to member organizations of the Wales Maritime andCoastal Partnership were interviewed and requested to participate.The researchers concluded that mapping stakeholders' values inthe marine environment was useful for identifying areas bettersuited for specific management regulations and for the develop-ment of comprehensive marine spatial plans.

There have been several non-marine participatory mappingstudies that have targeted stakeholders to assess protected areamanagement preferences. Eadens et al. (2009) conducted partici-patory mapping workshops with 35 individuals representing sixstakeholder groups for recreation planning in a Bahamian NationalPark. They modeled spatial agreement by examining the spatialoverlap in future activity zones mapped by the six groups. “Strong”agreement was defined as areas mapped by five to six groups and“some” agreement was defined as areas mapped by three to fourgroups. This method resulted in a park map showing areas ofspatial agreement for protection, ecotourism, and hunting activ-ities. In another example, Brown et al. (2015) used participatorymapping and non-spatial survey questions to identify public landvalues and preferences in the state of Victoria, Australia. Differentstakeholder groups were identified based on responses to surveyquestions asking about general preferences for public lands. Thesestakeholder groups were shown to have different place-specificpreferences depending on the public land type and location.

1.2. Stakeholder analysis for marine spatial planning

Spatial data collected using PPGIS/PGIS/VGI methods can beused in the early stages of planning to identify concentrations ofplace-specific marine values (both use and non-use values) thatwhen combined with ecological data, can identify preliminarymarine protected areas and/or management zones. Assessing thehuman dimensions of the marine environment through thisinductive, “bottom-up” approach presumes that high concentra-tions or “hotspots” of values will emerge from the participatorymapping activity. If the study area contains both existing andprospective MPAs, values mapped within existing MPAs can beused as an empirical basis for identifying similar areas for inclusionin the reserve system. The method was demonstrated by Raymondand Brown (2006) to identify the suitable areas for national parkexpansion in Victoria, Australia, based on the distribution of valueslocated in existing, proximate national parks. The supporting logicis that existing MPAs have place-specific values that differ fromsurrounding marine areas such that the type and relative abun-dance of these mapped values can be used to identify similarlyimportant areas.

As demonstrated in this paper, participatory mapping can alsobe used in the intermediate stages of a marine planning process toevaluate whether mapped values and preferences are consistentwith agency-proposed MPAs. A terrestrial analogue for thisapproach was a study by Brown (2006) on Kangaroo Island (KI),Australia that examined whether the type and distribution ofmapped values and preferences by KI residents were logicallyconsistent with development plan zones. This evaluative approachcan provide evidence in support of proposed MPAs or identify theneed for modification to MPA spatial design. When analyzing

mapped data within proposed MPAs, the potential for conflict be-tween specific stakeholder groups may become evident in thespatial distribution of mapped values and preferences. Examiningthe spatial distribution of mapped preferences appears moreimportant than values because preferences have a closer nexus tothe proposed purpose(s) for establishing an MPA.

The method for identifying stakeholder groups in participatorymapping is critically important because it determines how thespatial data are segmented and analyzed. Stakeholders can be pre-identified and recruited to engage in the participatory mappingactivity or non-spatial participant variables collected as part of themapping process can be used to identify stakeholder groups inpost-mapping analysis. Even if stakeholders are pre-identified, itwould appear prudent to also compare mapped values and pref-erences against presumed stakeholder roles.

The question of what constitutes a stakeholder for the purposeof marine spatial planning is non-trivial. The first complexity isjurisdiction. In the case of Australia, marine reserves can be createdin Commonwealth waters that extend from three nautical miles offthe coast to the outer limit of the exclusive economic zone (200nautical miles). Marine reserves established in coastal waters arethe responsibility of State governments. A second complexity isthat marine reserves for conservation function as quasi-publicgoods. A national or state government that designates a marinereserve for conservation may not be able to exclude others frombenefit, especially for pelagic species. Other nations, organizations,and individuals become stakeholders in the establishment andmanagement of MPAs. A third complexity influencing the delin-eation of stakeholders is the actual level of protection within MPAswhich can vary considerably from strict “no-take” zones to theallowance of extractive activities such as commercial fishing.

1.3. Stakeholder analysis of proposed MPAs in the Kimberley region,Australia

The aims of this study are to describe stakeholder analysismethods and to report findings from a participatory GIS process toassess coastal and marine values in the Kimberley region ofAustralia (Strickland-Munro, Kobryn, Moore, & Brown, 2015a).With limited research on stakeholder analysis methods usingparticipatorymapping, themethods assume equal importancewiththe actual results for marine planning in the study region. The stepsin stakeholder analysis and how each step was operationalizedappear in Fig. 1. We implemented the first three steps in this paperto inform a discussion about the fourth step, how and whether tointegrate stakeholder analysis into MPA decision support. Theoutput of the first step, stakeholder identification, influences sub-sequent steps in the process, emphasizing the importance of get-ting this step right. In this study, we used three operationalmethods for identifying stakeholder groups: participant self-identification with a group (identity), participant expression of in-terests in the study region (interests), and participant responses to atrade-off question that asked participants to express a preferencefor environmental or economic outcomes (value orientation). Thesestakeholder classifications formed the basis for the followingresearch questions:

(1) How should stakeholder groups be identified for purposes ofparticipatory mapping? We operationalize and evaluatethree methods for classifying participants into stakeholdergroups based on identity, interests, and general valueorientation.

(2) Are stakeholder groupings logically related to the type ofvalues and preferences mapped in participatory GIS? Weevaluate the propensity for different stakeholder groups to

Identify stakeholders

• Group identity• Interests• Value

orientation

Identify type and quantity of values and preferences by group

• General mapping behavior (measures of association; correpondence analysis)

Identify stakeholder preferences by place location

• Mapping behavior in MPAs or other areas of planning/mgmt interest (marker frequency distribution)

Place-based integration: aggregation and weighting of stakeholder preferences

• Weighted and unweighted aggregated responses by stakeholder group

Fig. 1. Sequence of steps in stakeholder analysis using participatory GIS for evaluating proposed MPAs.

G. Brown et al. / Applied Geography 67 (2016) 77e9380

map certain types of marine and coastal values andpreferences.

(3) Is stakeholder identity related to place-specific mappingbehavior? We evaluate the distribution of mapped prefer-ences by stakeholder groups in two of five proposed MPAs inthe study region using stakeholder definitions based onidentity and interests.

Following analyses and results, we reflect on the findings whichhave strong implications for the use of participatory mappingmethods for marine spatial planning. We provide some guidancefor participatory mapping processes that seek to integrate multiplestakeholder groups for decision support.

2. Methods

2.1. Study location and context

The Kimberley region is located in northwest Australia in thestate of Western Australia (see Fig. 2). The research study area ex-tends from the southwestern end of Eighty Mile Beach to theNorthern Territory border, a coastline 13,296 km in length at lowwater mark including islands. The marine environment of theKimberley is noted for its ‘very good’ ecological condition and isincluded in the 3.7% of global oceans considered to have experi-enced very low human impact (Halpern et al., 2008). In 2011, theWestern Australian Government introduced the Kimberley Scienceand Conservation Strategy (GoWA, 2011) with a commitment tointroduce a system of marine reserves through the establishment offour new, multiple-use marine parks located at Eighty Mile Beach,Roebuck Bay, Lalang-garram/Camden Sound and North Kimberley(see Fig. 2). The marine parks were to cover 48% of the Kimberley'scoastal waters and increase the area of State marine parks and re-serves from approximately 1.5 million hectares to 4.1 millionhectares (Thomson-Dans, Overman, & Moncrieff, 2011). A fifthmarine park for the iconic Horizontal Falls area was announced in2013 as well as plans to extend the North Kimberley Marine Parkeastwards to the Northern Territory border. To date, three parkshave been established, at Eighty Mile Beach, Horizontal Falls andLalang-garram/Camden Sound, with the remaining parks yet to beformalized. In Western Australia, marine parks include “no take”zones as well as “general use” zones where extractive activities areallowed. These existing and proposed State marine parks comple-ment four Commonwealth marine reserves located at Eighty MileBeach, Roebuck Bay, Argo-Rowley Terrace and ‘Kimberley’ (Fig. 2).Commonwealth marine reserves are managed primarily for biodi-versity conservation but also allow for a range of activitiesincluding commercial and recreational fishing, tourism, miningoperations, and pearling and aquaculture (CoA, 2014). All existingand proposed State marine parks are to be managed with Aborig-inal Traditional Owners under formal joint management

agreements.The principle economic activities associated with the Kimberley

coast include commercial fishing, pearling and other aquaculture(e.g., barramundi farming), oil and gas extraction, iron ore mining,and tourism. The Kimberley towns of Broome, Derby, Wyndhamand Kununurra are important service centers. The region's popu-lation is about 35,000 with 43.5% being of Aboriginal heritage (ABS,2011).

2.2. Data collection process

The research team designed, pre-tested and implemented aninternet-based PPGIS application for data collection. The applica-tion used a Google® maps interface where study participants coulddrag and drop digital markers onto a map of the Kimberley region(see Strickland-Munro et al., 2015a for a detailed description of thePPGISweb interface). The process consisted of participants enteringthe PPGIS website, providing informed consent, completing non-spatial survey questions (pre- and post-mapping), and engagingin the mapping activity. Pre-mapping questions included socio-demographic information, how respondents learned of the study,and their self-identified knowledge of the Kimberley region.

The post-mapping survey contained three questions designed toclassify participants into stakeholder categories based on groupidentity, interests, and value orientation. The first question askedparticipants to self-identify with a group based on the followingchoices: Kimberley resident; visitor; Aboriginal; commercial fish-ing, pearling or aquaculture; Commonwealth, state or local gov-ernment; NGO; tourism industry; oil/gas industry; and researcher.A second question asked participants regarding their dominantinterest in the Kimberley region. This question was framed byasking participants to indicate their greatest concern for the regionwith the following choices available: making sure there are recre-ational opportunities for local people; ensuring rights of TraditionalOwners/Aboriginal people in the region are respected; protectingbiological and ecological features found in the region; maintainingand developing tourism opportunities; ensuring the region pro-vides natural resources; and ensuring marine/coastal plans aredeveloped/implemented. A third question asked participants tothink about their own personal values and to position themselveson the 7-point Environmental-Economics Priority (EEP) scale,which contrasts environmental and economic priorities in coastaland marine management. Variants of the EEP scale have been usedin 19 studies indicating its reliability as a survey instrument(Abrams, Kelly, Shindler, and Wilton, 2005). In this study, the EEPwas used to classify participants into the categories of “environ-mental”, “balanced”, and “economic”. The scale was anchored atopposite ends with contrasting statements: “Highest priorityshould be given to maintaining natural environmental conditionseven if there are negative economic consequences” versus “Highestpriority should be given to economic considerations even if there

Fig. 2. Kimberley marine parks (current and proposed) (Source: Geoscience Australia 2014, Department of Parks and Wildlife).

G. Brown et al. / Applied Geography 67 (2016) 77e93 81

are negative environmental consequences.” The scalemidpoint wasanchored with the statement, “Environmental and economic fac-tors should be given equal priority.”

For the mapping activity, two different panels containedmarkers representing 14 values and 13 management preferences(see definitions in Table 1). The value markers were selected basedon inductive, emergent categories obtained from interview datacollected in an earlier phase of this research (see Strickland-Munro,Moore, Kobryn,& Palmer, 2015b), consideration of values unique tothe Kimberley region detailed in planning documents, and similarvalues found in a typology developed by Brown and Reed (2000)and used in multiple PPGIS studies (see Brown & Kytt€a, 2014).The selection of management preference markers was alsoinformed by these same interviews (e.g., key management issuesfor the region, see Strickland-Munro et al., 2015b), relevant policydocuments (e.g., Draft Kimberley Regional Planning and Infra-structure Framework, Government of Western Australia, 2014), andconsultation with key research partners including the WesternAustralia Marine Science Institute (WAMSI) and the WesternAustralia Department of Parks and Wildlife.

Sampling design and recruitment efforts were guided by thedesire to engage the greatest possible number of participants, a

formidable challenge given the Kimberley's vastness, small,dispersed population, and the region's limited accessibility. Thepopulation of interest included people living in or visiting theKimberley, as well as geographically-remote individuals with anongoing interest in the region. Stakeholder groups involved in aprior research phase were targeted for participation and includedAboriginal Traditional Owners; non-Aboriginal residents; touristsand the tourism industry; commercial and recreational fishing, andaquaculture; federal, state and local government; industry (mining,oil, gas and tidal energy interests); marine transport and aviation;and environmental non-government organizations. Sampling alsoincluded scientific researchers, particularly those involved in otherWAMSI research projects, and individuals from a commercial, on-line internet panel. A minimum target of 350 participants was setacross all stakeholder groups. In total, 120 official and informalrepresentative bodies were approached to participate in, and assistwith further recruitment for, the PPGIS survey over the months ofAprileJuly 2015.

Eight methods of recruitment were used to obtain PPGISparticipation: (1) direct personal contact by members of theresearch team, (2) postal invitation, (3) email initiated by stake-holder organizations that provided a link to the PPGIS website, (4)

Table 1Values and management preference markers with operational definitions.

Values Operational definition

Scenic/aesthetic These areas are valuable to me because they contain attractive scenery including sights, smells, and sounds.

Recreation These areas are valuable because they are where I enjoy spending my leisure time with family, friends or by myself, participating inoutdoor recreation activities (e.g., camping, walking, exploring).

Fishing (recreational) These areas are valuable because they are where I can go fishing for fish and other marine life like crabs, cockles, and oysters.

Economic (non-tourism) These areas are valuable because they provide natural resources that can be used by people (e.g., minerals, oil, gas, fish, pearls,pastoralism).

Nature-based tourism These areas are valuable because they provide tourism opportunities, including Aboriginal cultural tourism, in a generally undisturbedenvironment.

Learning/education/research

These areas are valuable because they enable us to learn about the environment through observation or study.

Biological/conservation These areas are valuable due to the presence of plants, wildlife & habitat including marine wildlife, reefs, migratory shorebirds &mangroves.

Aboriginal culture/heritage These areas are valuable because they allow Traditional Owners to maintain connection to their coastal & sea country through identityand place, family networks, spiritual practice and resource gathering.

European heritage These areas are valuable because they reflect European history associated with exploration, pastoralism, missions, commercial fishing &the Second World War.

Therapeutic/health These areas are valuable because they make me feel better mentally and/or physically.

Spiritual These areas are valuable because they are sacred, religious, or spiritually special places or because I feel reverence and respect for naturehere.

Intrinsic/existence These areas are valuable in their own right, no matter what I or others think about them.

Wilderness/pristine These areas are valuable because they are wild, uninhabited, or relatively untouched by European activity.

Special places These places are special. Please indicate why the place is special to you.

Preferences Operational definition

Increase conservation/protection

Increase conservation and protection here (e.g. from fishing pressure, encroaching development).

Increase aboriginalmanagement

Increase Aboriginal control and management of lands and waters, including ongoing resourcing for Ranger groups.

Add recreation facilities Add new recreation facilities (e.g. boat launching ramp, picnic area, campsite, toilet block).

Add tourism services/development

Add new nature-based tourism facilities (e.g., visitor center, eco-resort, pontoon).

Improve/increase access Improve or increase vehicular access (i.e., from no access to 4WD access or from 4WD track to 2WD road).

Restrict/limit access Restrict or limit access to protect environmental or culturally sensitive places, or to ensure the quality of visitor experiences.

Commercial fishing/aquaculture

Allow commercial fishing/aquaculture/pearling in this area.

No commercial fishing/aquaculture

Do not allow commercial fishing/aquaculture/pearling in this area.

Oil/gas development Allow oil/gas extraction and/or processing here.

No oil/gas development Do not allow oil/gas extraction and/or processing here.

New port development New port development here.

No new port development No new port development here.

Other preference Describe the land or sea use you would prefer (or not prefer) to see in this location.

G. Brown et al. / Applied Geography 67 (2016) 77e9382

social media, (5) local news media, (6) printed survey invitationcards, (7) announcement written in organization newsletters, and(8) informal referrals to friends, family, or professional contactsfrom any of the other methods.

A prototype of the PPGIS survey was pilot tested in March 2015using three approaches. The first approach requested differentgroups complete the survey, consisting of middle to senior levelmanagers in the WA Department of Parks and Wildlife, social sci-ence researchers at Murdoch University, and recreational users ofthe Kimberley coast. In the second approach, a member of theresearch team demonstrated the PPGIS survey in meetings withBroome-based participant groups. A third approach consisted of afocus group with individuals from the University of WesternAustralia. Feedback from these sources was used to adjust themapping scale, increase the clarity of mapping instructions, andadd extra place names and reference locations. The final version ofthe PPGIS survey was launched in April, 2015. Data were collected

for four months and the PPGIS survey was closed on July 31, 2015.

2.3. Analyses

2.3.1. Associations between stakeholder operational definitionsStakeholder groups were operationalized based on responses to

survey questions that asked participants about their identity, in-terests, and value orientation. There were nine stakeholder identitycategories, six interest categories, and three value orientation cate-gories. We examined the distribution of participants across thethree operational definitions using the chi-square test for inde-pendence to determine whether the alternative stakeholder clas-sifications were associated. Following a significant finding ofassociation, standardized residuals were calculated to assess whichpair-wise categorical variables most contribute to the overall as-sociation. The standardized residual was calculated by dividing theresidual value by the standard error of the residual. Standardized

G. Brown et al. / Applied Geography 67 (2016) 77e93 83

residuals greater than þ1.96 (rounded to þ2.0) indicated signifi-cantly greater observed frequencies than expected, while stan-dardized residuals less than �1.96 (rounded to �2.0) indicatedsignificantly fewer observed counts than expected. Larger absolutevalues of standardized residuals indicated greater deviation fromexpected counts, thus contributing more to the overall measure ofassociation.

2.3.2. Relationship between stakeholder group and mapped valuesand preferences (non-place specific)

To determinewhether stakeholder groups were logically relatedto the type of values and preferences mapped in the Kimberleyregion, we used chi-square analysis to determine if there was agreater propensity for different stakeholder groups to map certaintypes of values and preferences. If a participant mapped one ormore of a given value or preference marker category, that indi-vidual was classified as “YES” for the category, otherwise “NO”. Thiscategorical treatment of mapped markers (presence/absence) waspreferred over analyzing mean differences by stakeholder groupwhich can be influenced by a few individuals placing a largenumber of markers within a marker category.

We calculated chi-squared statistics and standardized residualsto determine whether the number of individuals within a stake-holder group mapping a given value or preference differed signif-icantly from the number of individuals that would be expected tomap the category. The chi-square analysis was supplemented withcorrespondence analysis to visualize the relationship betweenstakeholder groups and the types of values and preferences map-ped by the groups. Correspondence analysis describes the rela-tionship between two nominal variables in a contingency tablewhile simultaneously describing the relationships between thecategories of each variable. Mathematically, correspondence anal-ysis decomposes the chi-square measure of association of the twonominal variables into components, much like principal compo-nents analysis of continuous data. It computes row and columnscores and produces normalized plots based on the scores. In thenormalized plot, the distances between category points reflect therelationships between the nominal categories, with similar cate-gories plotted close to each other. Interpretation of the plot is byrows (i.e., stakeholder group) and columns (categories of values orpreferences).

2.3.3. Relationship between stakeholder group and place-specificpreferences for MPAs

To determine if stakeholder classification is related to place-specific mapping behavior, we examined the spatial distributionof mapped preferences by stakeholder group in two of the fiveproposed MPAs in the study regiondNorth Kimberley and RoebuckBay. These two areas were selected because i) the North Kimberleyproposed MPA had the greatest quantity of spatial data for analysisand is themost remote from human settlement, and ii) the RoebuckBay proposed MPA is the least remote with proximate humansettlement (Broome). We examined the spatial distributions in thetwo MPAs using the two operational definitions for stakeholdergroup, identity and interest. We generated radar (a.k.a., spider)charts of these preference frequency distributions for each stake-holder group to visually identify patterns of similarity and differ-ence for each MPA.

3. Results

3.1. Participation rates and response profile

A total of 763 individuals fully or partially participated in thePPGIS survey. A partial completion was an individual that accessed

the website and mapped one or more markers, but did not com-plete the post-mapping survey questions. Our analysis was limitedto full completions (n ¼ 578) because the stakeholder identityquestions were contained in the post-mapping survey questions. Ofthese participants, n ¼ 206 individuals originated from the onlineinternet panel while the remainder (n ¼ 372) came from otherrecruitment methods. Of all the recruitment methods, direct emailwas the most effective method, accounting for about 64% of par-ticipants. Social media and personal referral accounted for about13% and 8% of participants respectively. A postal mailing to Kim-berley residential households in the main population centers ofBroome, Derby, Wyndham and Kununurra (n ¼ 2915) was not aneffective recruitment strategy due, in part, to inaccurate postaladdresses, with about half of the letter invitations returned asundeliverable. Postal recruitment accounted for about 4% of par-ticipants with Kimberley residents accounting for approximately athird of study participants.

The sociodemographic profile of participants was examined andcompared to Kimberley and Western Australia census data (ABS,2011). Participants were 52% female compared to census data of50% for WA and 47% for the Kimberley region. The largest groups ofparticipants were aged 55e64 (21%), 35e44 (21%), and 45e54(20%) respectively, with this age profile being somewhat youngerthan comparable census data. Aboriginal participants were signif-icantly underrepresented in the response with only about 2% ofparticipants identifying themselves as Aboriginal compared to43.5% of the Kimberley population and the statewide proportion inWestern Australia of 3.4%. Participants were strongly biased towardhigher levels of formal education (bachelor or postgraduate de-grees), a finding consistent with previously reported PPGIS studies(Brown & Kytt€a, 2014).

3.2. Associations between stakeholder classifications

We generated chi-square contingency tables with standardizedresiduals to examine the distribution of participants across stake-holder classifications (identity, interests, and value orientation). Thelargest number of participants self-identified as visitors (n ¼ 271,51%) followed by government (n ¼ 68, 13%) and residents (n ¼ 61,11%). The smallest identity classification was commercial fishing(n ¼ 5, 1%). The largest stakeholder interest category was ecology(n ¼ 343, 60%) followed by Aboriginal (n ¼ 67, 12%), an interestingresult given that only 12 participants self-identified as Aboriginal.Stakeholder identity was significantly associated with stakeholderinterests (c2 ¼ 113.7, df ¼ 40, p < 0.001) with moderate strength ofassociation (Cramer's V ¼ 0.21, p < 0.001). There were multiple,significant pairwise associations (residuals > þ2.0) between oil/gasidentity and resource interests (þ3.8), Aboriginal identity andAboriginal interests (þ2.2), and resident identity with recreationalinterests (þ6.4) (see Table 2).

The largest number of participants selected an environmentalvalue orientation on the EEP scale (n ¼ 406, 71%), followed by abalanced orientation (n ¼ 116, 20%), and economic orientation(n ¼ 51, 9%). Stakeholder interest was significantly associated withvalue orientation (c2 ¼ 98.1, df ¼ 10, p < 0.001) with moderatestrength of association (Cramer's V ¼ 0.30, p < 0.001). The stan-dardized residuals indicate that participants with ecological in-terests were significantly over-represented in the environmentalvalue group (þ2.8), while recreation (�2.4), tourism (�2.0), andresource (�2.0) interests were under-represented (see Table 3). Theopposite relationships were found in the balanced group withrecreation (þ4.3), tourism (þ2.7), and resource (þ3.8) interestsover-represented, and ecology (�3.9) interests under-represented.In the economic group, ecology interests were also under-represented (�2.3).

Table 2Association of stakeholder identitywith stakeholder interest. The overall association is significant (c2 ¼ 113.720, df¼ 40, p < 0.001) with standardized residuals less than�2.0(pink) or greater than 2.0 (green) highlighted. Note: caution is warranted in interpreting results as 67% of cells have expected counts less than 5.

Identity InterestRecreation Aboriginal Ecology Tourism Resources Planning Total

Oil/gasCount 1 0 8 2 4 2 17% 5.9% 0.0% 47.1% 11.8% 23.5% 11.8% 100.0%Residual -.1 -1.4 -.8 .8 3.8 .4

TourismCount 1 1 16 4 0 3 25% 4.0% 4.0% 64.0% 16.0% 0.0% 12.0% 100.0%Residual -.5 -1.1 .2 1.8 -1.0 .5

GovernmentCount 4 7 46 2 1 4 64% 6.3% 10.9% 71.9% 3.1% 1.6% 6.3% 100.0%Residual -.2 -.2 1.1 -1.1 -1.1 -.8

NGOCount 0 3 18 2 0 1 24% 0.0% 12.5% 75.0% 8.3% 0.0% 4.2% 100.0%Residual -1.3 .1 .9 .3 -1.0 -.8

ResearchCount 0 8 37 0 0 2 47% 0.0% 17.0% 78.7% 0.0% 0.0% 4.3% 100.0%Residual -1.8 1.1 1.5 -1.8 -1.4 -1.1

AboriginalCount 1 4 6 0 1 0 12% 8.3% 33.3% 50.0% 0.0% 8.3% 0.0% 100.0%Residual .2 2.2 -.5 -.9 .7 -1.0

ResidentCount 17 2 27 3 5 6 60% 28.3% 3.3% 45.0% 5.0% 8.3% 10.0% 100.0%Residual 6.4 -1.9 -1.6 -.5 1.5 .2Count 11 34 156 20 10 28 259

Visitor % 4.2% 13.1% 60.2% 7.7% 3.9% 10.8% 100.0%Residual -1.6 .7 -.2 .7 -.3 .9

Commercial fishing

Count 0 1 1 1 1 1 5% 0.0% 20.0% 20.0% 20.0% 20.0% 20.0% 100.0%Residual -.6 .5 -1.2 1.2 1.7 .8

Total 35 60 315 34 22 47 5136.8% 11.7% 61.4% 6.6% 4.3% 9.2% 100.0%

Table 3Association of stakeholder interestwith value orientation. The overall association is significant (c2 ¼ 98.1, df¼ 10, p < 0.001) with standardized residuals less than�2.0 (pink)or greater than 2.0 (green) highlighted. Note: caution is warranted in interpreting results as 28% of cells have expected counts less than 5.

Value OrientationInterest

TotalRecreation Aboriginal Ecology Tourism Resources PlanningEnvironment Count 13 43 290 16 8 28 398

% 37.1% 65.2% 85.3% 44.4% 36.4% 54.9% 72.4%Std. Residual -2.4 -.7 2.8 -2.0 -2.0 -1.5

Balance Count 18 13 34 14 12 15 106% 51.4% 19.7% 10.0% 38.9% 54.5% 29.4% 19.3%Std. Residual 4.3 .1 -3.9 2.7 3.8 1.6

Economic Count 4 10 16 6 2 8 46% 11.4% 15.2% 4.7% 16.7% 9.1% 15.7% 8.4%Std. Residual .6 1.9 -2.3 1.7 .1 1.8

Total Count 35 66 340 36 22 51 550% 6.4% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0%

G. Brown et al. / Applied Geography 67 (2016) 77e9384

Stakeholder identity was significantly associated with value

Table 4Association of stakeholder identitywith value orientation. The overall association is signior greater than 2.0 (green) highlighted. Note: caution is warranted in interpreting result

Value orientation Oil/gas Tourism Government NGOEnvironment Count 7 20 54 2

% 41.2% 80.0% 85.7% 91.7Std. Residual -1.5 .5 1.3 1

Balance Count 8 3 8% 47.1% 12.0% 12.7% 8.3Std. Residual 2.5 -.9 -1.2 -1

Economics Count 2 2 1% 11.8% 8.0% 1.6% 0.0Std. Residual .4 -.1 -1.9 -1

09Total Count 17 25 63 2% 3.3% 100.0% 100.0% 100.0

orientation (c2 ¼ 52.9, df¼ 16, p < 0.001) withmoderate strength of

ficant (c2¼ 52.9, df¼ 16, p < 0.001) with standardized residuals less than�2.0 (pink)s as 44% of cells have expected counts less than 5.

Identity

TotalResearch Aboriginal Resident VisitorCommercial

fishing2 40 8 34 178 2 365

% 88.9% 66.7% 56.7% 69.0% 40.0% 71.7%.2 1.4 -.2 -1.4 -.5 -.82 4 0 19 53 3 100

% 8.9% 0.0% 31.7% 20.5% 60.0% 19.6%.3 -1.6 -1.5 2.1 .3 2.00 1 4 7 27 0 44

% 2.2% 33.3% 11.7% 10.5% 0.0% 8.6%.4 -1.5 2.9 .8 1.0 -.74 45 12 60 258 5 509

% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0%

G. Brown et al. / Applied Geography 67 (2016) 77e93 85

association (Cramer's V ¼ 0.23, p < 0.001) (see Table 4). The stan-dardized residuals indicate that participants with oil/gas (þ2.5),resident (þ2.1), and commercial fishing (þ2.0) interests weresignificantly over-represented in the balanced value group, whileAboriginal (þ2.4) interests were over-represented in the economicgroup.

3.3. Stakeholder values and preferences (non-place specific)

The non-place specific mapping behavior of stakeholder groupsbased on identity, interests, and value orientation were analyzedusing chi-square and correspondence analyses. With respect toself-identified group, residents (n ¼ 120) were more likely to maprecreation values and preferences to increase recreation facilities,tourism stakeholders (n ¼ 33) were more likely to map nature-based tourism values and preferences to limit oil/gas develop-ment, NGOs (n ¼ 26) were more likely to map biological/conser-vation values and preferences to increase conservation protection,and government (n ¼ 86) and research (n ¼ 66) stakeholders weremore likely to map biological conservation values and preferencesto limit new port development (see Table 5). There were relativelyfew participants that self-identified with the commercial fishing(n ¼ 5) and oil/gas industries (n ¼ 19), but the mapped preferencesof these participants were consistent with these identities, withcommercial fishing stakeholders more likely to map preferences toincrease commercial fishing and oil/gas stakeholders more likely tomap preferences to increase oil/gas development. Stakeholdersidentifying as Aboriginal (n ¼ 12) were less likely to map prefer-ences to add tourism facilities and less likely to prohibit oil/gasdevelopment. Visitors (n ¼ 343) were more likely to map recrea-tional fishing values and less likely to map preferences for new portdevelopment.

Stakeholder interests that were significantly related to type andnumber of mapped values and preferences appear in Table 6.

Table 5Stakeholder groups by identity that are significantly related (p� 0.05 yellow or p� 0.10 gsquare test of independence was calculated for each stakeholder group in a 2 � 2 contindividuals in the stakeholder group mapped than individuals not in the group while minP-values of the chi-square association are also reported.

Stakeholder identity ValuesResident (n=120) Recreation (+) .001

Rec. fishing (+) .000

Aboriginal (n=12)

Visitor (n=343) Rec. fishing (-) .001

Commercial fishing (n=5) Recreation (-) .074Nature-based tourism (-)Wilderness (-) .010

Government employee (n=86)

Rec. fishing (+) .042Biological/conservation (European heritage (+) .0

NGO (n=26) Scenic (+) .039Biological/conservation (Wilderness (+) .10Special places (+) .070

Tourism (n=33) Nature-based tourism (+Special places (+) .10

Oil/gas (n=19) Aboriginal culture (-) .096

Research (n=66) Recreation (-) .000Rec. fishing (-) .000Economic (non-tourism) Biological/conservation (Therapeutic/health (-) .00

Stakeholders with recreation interests (n ¼ 35) were more likely tomap fewer nature-based tourism (�2.8), intrinsic (�2.1), Aboriginalculture (�3.2), biological (�4.1), and wilderness values (�4.4), andsignificantly more preferences to improve access (þ2.2). Stake-holders identifying with Aboriginal interests were more likely tomap fewer wilderness values (�2.1), while stakeholders withecological interests were more likely to map intrinsic (þ3.0),learning/research (þ3.4), nature-based tourism (þ3.4), Aboriginalculture (þ4.1), biological (þ6.1), and wilderness values (þ6.3).Ecological interests were more likely to map preferences to limitnew oil/gas (þ3.1) and port development (þ2.9), and to increaseconservation (þ5.5). Tourism interests were more likely to mappreferences to add recreation facilities (þ2.5) and improve access(þ2.0), while resource interests were more likely to map fewerpreferences to increase conservation (�2.3), increase Aboriginalmanagement (�2.3), and to limit oil/gas (�3.0) and port develop-ment (�2.1). Planning stakeholder interests were more likely tomap values for biological (þ2.2) and recreational fishing (þ2.0).

Stakeholder groups by value orientation were unevenlydistributed between environmental (n ¼ 406), balanced (n ¼ 116),and economic (n ¼ 51) priorities (see Table 7). Environmentalstakeholders were more likely to map values of most types, espe-cially biological (þ6.9) and wilderness (þ6.1) values, while thebalanced and economic groups mapped fewer of the same cate-gories of values. Stakeholder group propensities tomap preferenceswere logically related to the types of values that were mapped. Forexample, environmental stakeholders were more likely to mappreferences to restrict access, limit commercial fishing, and restrictnew oil/gas and port development. The balanced group expressedthe opposite pattern for mapped preferences. These results suggesta stronger pro-development perspective in the balanced group thanthe environmental group. The balanced group also appearedsomewhat more pro-development than the economic group, aresult inconsistent with what would be expected on the EEP scale.

reen) to the type and number of values and management preferences mapped. A chi-ingency table: group/non-group by mapped/non-mapped. Plus (þ) indicates moreus (�) indicates fewer individuals in the group mapped than others not in the group.

PreferencesAdd recreation facilities (+) .071No commercial fishing (+) .009New port development (+) .000Other preference (+) .002Add tourism services (-) .078No oil/gas development (-) .070No oil/gas development (-) .089New port development (-) .041Other preference (-) .035

.043Commercial fishing (+) .021

+) .00777

No new port development (+) .091Other preference (+) .074

+) .025Increase conservation (+) .001Increase Aboriginal management (+) .036

) .085 No oil/gas development (+) .012

Oil/gas development (+) .003

(-) .093+) .0004

Restrict access (+) .085No new port development (+) .026

Table 6Stakeholder groups by interest that are significantly related to the type and number of values and management preferences mapped. Numbers in parentheses indicatesignificant standardized residuals (greater than þ2.0 or less than �2.0) following a statistically significant chi-square association (p � 0.10) between the six stakeholderinterests and a given value or preference. The sign of the standardized residual indicates more (green þ) or fewer (yellow �) individuals in the stakeholder interest mappedthan expected.

Stakeholder interest Values PreferencesRecreation (n=35) Nature-based tourism (-2.8)

Biological/conservation (-4.1)Aboriginal culture (-3.2)Intrinsic (-2.1)Wilderness (-4.4)

Increased conservation (-3.1)Improve access (+2.2)

Aboriginal (n=67) Wilderness (-2.1)Ecological (n=343) Nature-based tourism (+3.4)

Biological/conservation (+6.1)Aboriginal culture (+4.1)Education/research (+3.4)Intrinsic (+3.0)Wilderness (+6.3)Special places (+3.1)

Increased conservation (+5.5)No oil/gas development (+3.1)New port development (-2.2)No new port development (+2.9)

Tourism (n=36) Biological/conservation (-3.8)Aboriginal culture (-2.3)Wilderness (-2.5)

Add recreation facilities (+2.5)Improve access (+2.0)

Resources (n=22) Biological/conservation (-2.7)Education/research (-2.5)Wilderness (-2.4)

Increased conservation (-2.3)Increased Aboriginal management (-2.3)No oil/gas development (-3.0)No new port development (-2.1)

Planning (n=51) Biological/conservation (+2.2)Rec. fishing (+2.0)

Table 7Stakeholder groups by value orientation (environment/balanced/economic) that are significantly related to the type and number of values and management preferencesmapped. Numbers in parentheses show significant standardized residuals (greater than þ2.0 or less than �2.0) following a statistically significant chi-square association(p � 0.10) between the three stakeholder orientations and a given value/preference. The sign of the standardized residual indicates more (green þ) or fewer (yellow �) in-dividuals in the stakeholder orientation mapped than expected.

Stakeholder value orientation

Values Preferences

Environmental (n=406) Scenic (+3.5) Nature-based tourism (+3.4)Biological/conservation (+6.9)Aboriginal culture (+4.6)European heritage (+3.5)Intrinsic (+3.2)Wilderness (+6.1)

Increase conservation (+4.9)Increase Aboriginal mgmt. (+3.2)Add recreation facilities (-2.7)Restrict access (+2.2)Commercial fishing (-2.3)No commercial fishing (+2.5)Oil/gas development (-4.3)No oil/gas development (+5.4)New port development (-3.9)No new port development (+3.7)

Balanced (n=116) Scenic (-2.2)Nature-based tourism (-2.7)Biological/conservation (-5.8)Aboriginal culture (-5.1)European heritage (-3.1)Intrinsic (-3.2)Wilderness (-5.4)

Increase conservation (-3.4)Increase Aboriginal management (-2.3)Add recreation facilities (+2.3)Restrict access (-2.3)Commercial fishing (+2.6)No commercial fishing (-2.0)Oil/gas development (+3.5)No oil/gas development (-4.8)New port development (+3.6)No new port development (-3.1)

Economic (n=51) Scenic (-2.5)Biological/conservation (-2.9)Wilderness (-2.0)

Increase conservation (-3.0)

G. Brown et al. / Applied Geography 67 (2016) 77e9386

Correspondence analysis was used to generate normalized plotsof the relationships between stakeholder groups defined by identityand interests and the categories of mapped values and preferences.Correspondence analyses of stakeholder groups by mapped pref-erences captured more of the total inertia or variance explained(identity ¼ 20%, interests ¼ 18%) than mapped values(identity ¼ 12%, interests ¼ 12%). With stakeholder identity, NGOsand research stakeholders were similar in their propensity to mapbiological/conservation values, Aboriginal stakeholders and resi-dents were similar in mapping recreational fishing and specialplace values, and tourism and government stakeholders weresimilar in their propensity to map scenic, Aboriginal culture, naturetourism, and learning values (Fig. 3). Research, tourism, and

Aboriginal stakeholders were similar in their propensity to mappreferences to restrict oil/gas and new port development; visitorshad greater propensity to map preferences to increase recreationfacilities, tourism development, access, and Aboriginal manage-ment; and NGOs had greater propensity to map preferences forincreasing conservation and Aboriginal management. Commercialfisherman and oil/gas stakeholders were differentiated from theother stakeholder groups in their propensity tomap preferences forincreasing commercial fishing and new oil/gas developmentrespectively.

The normalized plots of stakeholder groups by interests areprovided in Fig. 4. Stakeholders with interests in resources andrecreation were differentiated from other interests in their

Fig. 3. Correspondence analysis plots by self-identified stakeholder group by mapped: (a) place values and (b) management preferences.

Fig. 4. Correspondence analysis plot of stakeholder groups by self-identified future interests with mapped: (a) place values and (b) management preferences.

G. Brown et al. / Applied Geography 67 (2016) 77e93 87

propensity to map recreational fishing value, while planning in-terests were most closely associated with the mapping of scenic,economic (non-tourism), and recreation values. Ecological interestswere associated with the mapping of wilderness, biological,intrinsic, spiritual, Aboriginal culture, and special place values.With respect to mapped preferences, there were clearer associa-tions by stakeholder interests. Resource interests were associatedwith the mapping of preferences for new oil/gas and port devel-opment, recreation interests were associated with the mapping of

increased access and commercial fishing, Aboriginal interests wereassociated with mapped preferences to increase Aboriginal man-agement in the region, tourism interests were associated withpreferences to increase tourism development and recreation facil-ities, and ecological interests were associated with preferences tolimit access and all types of development while increasingconservation.

G. Brown et al. / Applied Geography 67 (2016) 77e9388

3.4. Stakeholder preferences by MPA

The frequency distribution of mapped preferences for all studyparticipants for the five proposed MPAs in the region appears inFig. 5. The number of preferences ranged from n ¼ 1133 in theproposed North Kimberley MPA to n ¼ 155 in the proposed EightyMile BeachMPA. The distribution of profiles was very similar for thenorthern MPAs (North Kimberley, Horizontal Falls, and CamdenSound/Lalang Garram) with the largest number of preferences forincreased conservation and restricting commercial fishing and oil/gas. The southern MPAs (Roebuck Bay and Eighty Mile Beach) hadsomewhat greater proportions of preferences, relative to thenorthern Kimberly MPAs, of increasing recreation facilities andaccess.

The frequency distributions of preferences were generated forthe North Kimberley and Roebuck Bay proposed MPAs using thestakeholder classifications for identity and interests (Figs. 6 and 7).The frequency distributions of preferences for the North Kimberleyproposed MPA by identity indicate that NGO, tourism, and visitorstakeholders strongly emphasized conservation preferences in thisarea. Research stakeholders placed greater emphasis on Aboriginalmanagement and decreased access than other stakeholder groups.The mapping of preferences in the North Kimberley by oil/gasstakeholders emphasizing the restriction of new oil/gas and portdevelopment may appear counter-intuitive. However, these resultswarrant caution in interpretation given that mapped preferencedata for oil/gas stakeholders was quite limited overall. Further, it isnoteworthy that the largest percentage of participants by oil/gasidentity identified with ecology interests (see Table 2).

In the North Kimberley, there were some differences in therelative proportions of preferences mapped by stakeholder in-terests. Participants identifying with resources placed greateremphasis on new port development while participants identifyingwith Aboriginal interests placed greater emphasis on Aboriginalmanagement. Stakeholder interests associated with planning,tourism, and recreation placed greater emphasis on increasing rec-reation facilities than the other groups (see Fig. 6). In the RoebuckBay proposed MPA, preferences for conservation were dominantamong stakeholder interests with tourism and recreation interests

Fig. 5. Mapped management preferences by category (%) for all participants in proposed maHorizontal Falls (c) Camden Sound (d) Roebuck Bay (e) Eighty Mile Beach.

showing stronger preferences for increased access (see Fig. 7).The frequency distributions of preferences for the two proposed

MPAs by identity indicate that NGO and tourism stakeholdersstrongly emphasized conservation preferences in the North Kim-berley while residents strongly emphasized conservation prefer-ences in Roebuck Bay. (see Figs. 8 and 9). Research stakeholdersplaced greater emphasis on Aboriginal management and decreasedaccess in both proposed MPAs than other stakeholder groups. InRoebuck Bay, visitors placed greater emphasis on increased recre-ation facilities and access. The mapping of preferences in the NorthKimberley by oil/gas stakeholders that emphasized the restrictionof new oil/gas and port development may appear counter-intuitive.However, these results warrant caution in interpretation given thatmapped preference data for oil/gas stakeholders was quite limitedoverall with no mapped preferences in Roebuck Bay.

4. Discussion

We have presented a method for conducting stakeholder anal-ysis for marine conservation planning by operationalizing multipledefinitions for stakeholder groups based on identity, interests, andvalue orientation. These stakeholder classifications were signifi-cantly associated, with stakeholder identity logically related tostakeholder interests in the study region. Stakeholder groupsshowed significantly greater propensity for mapping certain typesof values and preferences related to their identity, interests, or valueorientation. Analysis of mapping behavior of stakeholder groups inproposed MPAs in the Kimberley region revealed that managementpreferences can be differentiated based on stakeholder identity orinterests. The implications for marine conservation and futurePPGIS efforts are explored in the following two sections.

4.1. Marine conservation planning implications

The operational method for identifying stakeholder groups af-fects what values and preferences are mapped. Although we founda statistical association between self-identified stakeholder identityand interests, the association was not strong, resulting in mappingdifferences between identity and interests. For example,

rine parks in the Kimberley region: Clockwise from left to right: (a) North Kimberley (b)

Fig. 6. Mapped management preferences by category (%) in the proposed North Kimberley Marine Park by stakeholder interest groups. Clockwise from left to right: (a) all groups(b) ecology (c) resources (d) Aboriginal (e) planning (f) tourism (g) recreation.

Fig. 7. Mapped management preferences by category (%) in the proposed Roebuck Bay Marine Park by stakeholder interest groups. Clockwise from left to right: (a) ecology (b)Aboriginal (c) planning (d) tourism (e) recreation.

G. Brown et al. / Applied Geography 67 (2016) 77e93 89

participants that identify with the tourism industry dispropor-tionately mapped preferences in opposition to oil/gas develop-ment, but participants expressing future interests in tourismmapped disproportionately more preferences for adding recreationfacilities and increasing access. This occurred because only a smallproportion (16%) of participants that self-identified with tourismindicated their primary future interest in the region to be tourism(see Table 2). The majority of tourism stakeholders by identitywerenot the same individuals expressing tourism interests. Thesestakeholder classification differences can manifest in differentplanning priorities for proposed MPAs. Tourism stakeholders byidentity and interests mapped similar priorities for the North Kim-berley MPA, but the Roebuck Bay MPA mapping outcomes weredifferent, with tourism identity stakeholders prioritizing increasedconservation, and tourism interest stakeholders prioritizing

increased access. From a marine planning perspective, these dif-ferences may be important depending onwhether increased accessis considered compatible or incompatible with increasedconservation.

An important, but under-researched topic in marine conserva-tion planning is the influence of agency planners, policy makers,and managers on planning outcomes. The majority of participantsthat identified with the government stakeholder group expressedtheir future interests in the Kimberley as ecological (Table 2, 72%),with few individuals (n ¼ 4) indicating their future interests in theKimberley to be planning. The largest stakeholder group by identitythat expressed interest in marine planning was visitors, but thesevisitors are unlikely to hold formal positions with direct influenceover MPA planning outcomes. In this case, the implications formarine planning are best viewed through participants having both

Fig. 8. Mapped management preferences by category (%) in the proposed North Kimberley Marine Park by stakeholder identity groups. Clockwise from left to right: (a) oil/gas (b)tourism (c) government (d) NGO (e) research (f) resident (g) visitor.

Fig. 9. Mapped management preferences by category (%) in the proposed Roebuck Bay Marine Park by stakeholder identity groups. Clockwise from left to right: (a) tourism (b)government (c) NGO (d) research (e) resident (f) visitor.

G. Brown et al. / Applied Geography 67 (2016) 77e9390

government identity and as well as planning interests, but the smallnumber of participants meeting these criteria (n ¼ 4) exposes alimitation of the methods described herein. These four individualsmay be highly influential in MPA planning within government, butthere is no way to determine their identity or relative decision in-fluence. Further, spatial mapping with points, as was done in thisstudy, requires relatively large sample sizes to make valid in-ferences about spatial locations. With smaller sample sizes, the useof polygons for spatial mappingmay bemore appropriate (Brown&Pullar, 2012). Thus, while the small number of stakeholders thatidentify with both government and planning interests appear sup-portive of conservationwith opposition to resource development inthe two proposed MPAs examined, these results require more in-formation about the participants for meaningful interpretation.

In marine conservation planning, recreation and ecology

interests are often assumed as having little in common. However,this assumption is usually madewith little or no supporting data. Inthis study, there was empirical evidence showing different mappedvalues and preferences by recreation and ecology interests (seeTable 6), resulting in different planning priorities in the two pro-posed MPAs (see Figs. 6 and 7). Conservation interests were mostconcerned with restricting resource development activities whilerecreation interests were most concerned with increasing access.Understanding these differences early in the marine planningprocess is essential to address latent interests that can thwarteffective planning outcomes.

Analyzing stakeholder mapping data at different geographicscales and levels of data aggregation provides different insights formarine conservation planning. While our final analysis focused onspecific, proposed MPAs in the study region, there is also value in

G. Brown et al. / Applied Geography 67 (2016) 77e93 91

examining whole region and aggregated results. When the mappedpreferences for all participants were combined, preferences formarine conservation dominated as illustrated in Fig. 5. We surmisethis result reflects strong engagement by conservation interests inthis study (i.e., 71% of participants expressed an environmentalvalue orientation). This level of engagement may reflect the highlypoliticized process of marine conservation planning in the Kim-berley region where conservation efforts have recently emergedand galvanized to oppose a proposal for major offshore gas devel-opment and associated land-based infrastructure. Thus, the timingof a participatory mapping process relative to important externalevents can strongly influence the type of people that participateand thus, the mapped results.

A comparative analysis of management preferences across thefive MPAs using radar plots (Fig. 5) demonstrates the usefulness ofsuch plots for identifying potential regional differences in desiredplanning outcomes. For the two more accessible MPAs close toBroome (Roebuck Bay and Eighty Mile Beach), increasing accessand recreation facilities were more favored than for the three moreremote MPAs. This comparative analysis provides a means formarine planners to identify and potentially tailor MPA manage-ment options based on regional differences. MPAs located nearlarger human settlements will likely require greater flexibility inmanaging the MPAs to accommodate a wider range of human uses.

4.2. Stakeholder analyses and future PPGIS efforts

A facile conclusion of our findings would be that stakeholdergroup affiliation, whether through identity or interests, determineswhat values and preferences people are likely to map in partici-patory GIS. For example, NGOs representing environmental valueswere strongly supportive of conservationwhile restricting resourcedevelopment, commercial fisherman favored commercial fishingactivity, and oil/gas stakeholders favored oil/gas development.However, the data and conclusions from this study appear morenuanced. For example, stakeholders identifying with the oil/gasindustry expressed less value for Aboriginal culture, researchstakeholders expressed less value for recreational and therapeuticvalues that derive from increased access to remote areas, and vis-itors expressed less value for recreational fishing despite beingmarketed as a regional attraction.

The identification of stakeholder identity using value orientationprovided limited explanatory power, the likely result of socialdesirability bias in participant EEP scale responses. Stakeholderswith resource development interests were more likely to select abalanced value orientation over economic prioritization. The otherstakeholder operational definitions, identity and interests, providedbetter insight into mapped preferences for marine conservation.There was greater propensity for participants that identified orexpressed interests in resources to map more development pref-erences for oil/gas and port development, and these general map-ping propensities were evident in place-specific, proposed MPAs.Participants identifying with environmental NGOs or expressingecological interests had greater propensity to map conservationand anti-resource development preferences which also manifestedin proposed MPAs. These results support the findings of Brown(2013) that non-spatial values of participants can manifest inbehavioral choices when mapping place-specific values and pref-erences, and that volunteer sampling, in particular, can result inbiased perspectives toward resource use and environmental pro-tection compared to random sampling methods (Brown, Kelly, &Whitall, 2014). Knowing which particular stakeholder groups andinterests participated in the mapping activity appears to be criticalinformation for determining how to aggregate and potentiallyweight responses for determining the acceptability of new

proposed MPAs.We described the final step in stakeholder analysis using

participatory mapping as place-based integration for decisionsupport. But what does this mean? As observed byWeible (2007) inthe context of MPAs, this means going beyond technical analysisand engaging in political analysis. A common stakeholder analysismethod identifies and maps stakeholders in two-dimensionalspace consisting of power/influence by level of interest (Bryson,2004). However, stakeholder analysis using PPGIS does notexplicitly assess stakeholder power or influence. Further, whereastraditional stakeholder analysis usually focuses on a single alter-native (Ramirez, 1999; Susskind & Thomas-Larmer, 1999), partici-patory mapping provides stakeholder information on multiplepreferences and futures, and in this study, multiple MPAs. Volun-teer sampling, as undertaken in this study, can result in biasedperspectives toward resource use and environmental protectioncompared to random sampling methods (Brown et al., 2014;Brown, in press).

Brown et al. (2015) describedmethods formodeling stakeholderagreement and disagreement in multiple, place-specific locationsto inform management decisions and emphasized the need forresearch that provides critical insight into stakeholder dynamics,power relations, and perceptions of influence over governmentofficials responsible for decision-making. The integration of suchinformation with participatory mapped spatial data is needed toprogress the utility of PPGIS in decision support. Without this in-formation, the weightings idealized in Fig. 1 can only be providedthrough speculation about power relations that may not reflectreality.

A potential approach to the vexing issue of stakeholderweighting would be to conduct a representative survey of Austra-lian residents to determine how their values and preferences align(or not) with the different stakeholder perspectives identified inthis study. The results of the survey could be used to derivestakeholder weights for aggregating preferences for proposedMPAs in the region. However, one important limitation with thisapproach is the different ontological assumptions underpinningsuch a survey and PPGIS. A general survey is likely to evoke re-sponses that are place-independent, eliciting responses from in-dividuals potentially unfamiliar with the marine and coastalenvironments of the study region. In contrast, PPGIS is under-pinned by the ontological assumption that participants areproviding responses based on some place familiarity.

In the absence of political analysis or a survey of Australianresidents, the default position for stakeholder analysis usingparticipatory mapping methods assumes that stakeholders aresimilar in importance and influence, and accordingly, their mappedvalues and preferences can be aggregated and interpreted withoutweighting. Wewould consider this approach politically naïve giventhe highly contentious nature of marine conservation in Australiaand elsewhere. In the case of the Kimberley region, the aggregated,unweighted responses of stakeholder groups do not suggest strongconflict over the proposed MPAs. The coastal and marine values, aswell as management preferences, were strongly supportive ofconservation as a priority. However, commercial fishing interestswere largely absent from the participatory mapping process wherethey often present the most vocal opposition to MPAs, although thelevel of oppositionwill depend on the actual zoning of the MPA andwhether the MPA is designated as a “no-take” area. The Kimberleyregion does have offshore oil/gas development potential, but thenear-shore location of the proposed MPAs in state waters make oil/gas development less politically feasible in the region, with themapped results reflecting this current political position.

Despite efforts to engage resource development interests in theparticipatory mapping process, the largely volunteer sampling and

G. Brown et al. / Applied Geography 67 (2016) 77e9392

recruitment resulted in participants that were demographicallybiased toward younger and more highly educated individuals, andimportantly, toward mapped values and preferences that favorcoastal and marine conservation. The extent of influence of thislatter bias is unknown. The level of engagement by commercialfishing (n ¼ 5) and oil/gas (n ¼ 19) interests in this study was lowrelative to expectations regarding engagement by these stake-holders in other jurisdictions in Australia. However, the absence ofcommercial fishing interests from the participatory mapping pro-cess probably reflected the relatively small number of commercialfishing operations in the region relative to the eastern seaboard ofAustralia where such interests have been vocal opponents to MPAdesignation. And, although the Kimberley region has considerableoil/gas reserves, their offshore location relative to the nearshorelocation of the proposed MPAs in State waters means any contro-versy over this activity is less likely to be reflected in this study.

Whether stakeholders are “latent” or “definitive” (likely to seekpolitical influence on planning outcomes) (Mitchell et al., 1997)matters in terms of possible weightings of values and preferencesin PPGIS. In this study, the largest group of stakeholders by identityconsisted of visitors who are best described as latent stakeholders.More definitive stakeholders such as commercial fishing and oil/gasinterests are likely to demonstrate political influence far beyondwhat their participatory mapping engagement in this study wouldsuggest.

5. Conclusion

Crowd-sourcing methods using PPGIS can result in betterplanning decisions (Brown, 2015) with the quality of these de-cisions enhanced by understanding how stakeholder identificationmethods influence interpretation of the results. Even knowingwhich particular stakeholder groups and interests participated inthe mapping activity is critical information for determining how toaggregate information on proposed MPAs. The value of stakeholderidentification using value orientation is questionable given it pro-vided limited explanatory power, a likely result of social desirabilitybias in participant EEP scale responses.

Two major challenges require research and practitioner atten-tion. The first is how to ensure a full suite of stakeholder engage-ment beyond conservation interests, which is essential if PPGIS is tobe a credible, defensible approach to social data collection andanalysis. Second, and probably more important, is being able toweight or use other means to include consideration of the relativepower of stakeholders, their perceptions, and the perceptions ofthose involved in, and leading decision making (e.g., elected offi-cials, government employees).

This is one of the first studies to report on the use of partici-patory mapping methods (PPGIS/PGIS/VGI) with the aim ofinforming marine spatial planning. This is not surprising given thatparticipatory mapping to inform terrestrial conservation is rela-tively recent as well. The application of PPGISmethods is a responseto increasing calls for social research to inform MPA planning andmanagement (Gruby et al., 2015; Voyer et al., 2012). As these au-thors attest, the barriers to the inclusion of social science, includingPPGIS, are more political than technical. Voyer et al. (2012) note thecontinuing barriers to effective public participation, while Grubyet al. (2015) make a more fundamental plea regarding inclusionof social dimensions in research for large MPAs. As part of a largemultidisciplinary research program designed to inform manage-ment, this study has the potential to contribute significantly to theplanning for Kimberley MPAs. As such, the story for the Kimberleycoastal region is still being written.

Acknowledgments

The time and expertise contributed by the people who partici-pated in this project are acknowledged. Without their generosity,this research would not have been possible. The interest and sup-port of the Department of Parks and Wildlife and the 120 in-dividuals, formal and informal groups who were involved inrecruitment are also gratefully acknowledged. This research wassupported by the Kimberley Marine Research Program, adminis-tered by the Western Australian Marine Science Institution.

References

Abrams, J., Kelly, E., Shindler, B., & Wilton, J. (2005). Value orientation and forestmanagement: the forest health debate. Environmental Management, 36(4),495e505.

ABS. (2011). 2011 census data (Accessed 27 August) http://www.censusdata.abs.gov.au.

Australia Department of Environment. (2015). Commonwealth marine reservesoverview. Retrieved from http://www.environment.gov.au/topics/marine/marine-reserves/overview.

Brown, G. (2005). Mapping spatial attributes in survey research for natural resourcemanagement: methods and applications. Society & Natural Resources, 18(1),1e23.

Brown, G. (2006). Mapping landscape values and development preferences: amethod for tourism and residential development planning. International Journalof Tourism Research, 8(2), 101e113.

Brown, G. (2012). Public participation GIS (PPGIS) for regional and environmentalplanning: reflections on a decade of empirical research. URISA Journal, 25(2),5e16.

Brown, G. (2013). Relationships between spatial and non-spatial preferences andplace-based values in national forests. Applied Geography, 44, 1e11.

Brown, G. (2015a). Engaging the wisdom of crowds and public judgment for landuse planning using public participation GIS (PPGIS). Australian Planner, 52(3),199e209.

Brown, G. (2015b). A review of sampling effects and response bias in internetparticipatory mapping (PPGIS/PGIS/VGI). Transactions in GIS. In press.

Brown, G., de Bie, K., & Weber, D. (2015). Identifying public land stakeholder per-spectives for implementing place-based land management. Landscape and Ur-ban Planning, 139, 1e15.

Brown, G., Kelly, M., & Whitall, D. (2014). Which “public”? Sampling effects in publicparticipation GIS (PPGIS) and volunteered geographic information (VGI) sys-tems for public lands management. Journal of Environmental Planning andManagement, 57(2), 190e214.

Brown, G., & Kytt€a, M. (2014). Key issues and research priorities for public partici-pation GIS (PPGIS): a synthesis based on empirical research. Applied Geography,46, 122e136.

Brown, G., & Pullar, D. (2012). An evaluation of the use of points versus polygons inpublic participation geographic information systems (PPGIS) using quasi-experimental design and Monte Carlo simulation. International Journal ofGeographical Information Science, 26(2), 231e246.

Brown, G., & Reed, P. (2000). Validation of a forest values typology for use in na-tional forest planning. Forest Science, 46(2), 240e247.

Bryson, J. M. (2004). What to do when stakeholders matter: stakeholder identifi-cation and analysis techniques. Public Management Review, 6(1), 21e53.

Charles, A., & Wilson, L. (2009). Human dimensions of marine protected areas. ICESJournal of Marine Science: Journal du Conseil, 66(1), 6e15.

Christie, P. (2004). Marine protected areas as biological successes and social failuresin Southeast Asia. In American fisheries society symposium (Vol. 42, pp.155e164).

Christie, P., McCay, B. J., Miller, M. L., Lowe, C., White, A. T., Stoffle, R., et al. (2003).Toward developing a complete understanding: a social science research agendafor marine protected areas. Fisheries, 28(12), 22e25.

Chuenpagdee, R., Pascual-Fern�andez, J. J., Szeli�anszky, E., Alegret, J. L., Fraga, J., &Jentoft, S. (2013). Marine protected areas: re-thinking their inception. MarinePolicy, 39, 234e240.

Commonwealth of Australia. (2014). North-west commonwealth marine reservesnetwork (Accessed 2 July) http://www.environment.gov.au/topics/marine/marine-reserves/north-west/management.

Eadens, L. M., Jacobson, S. K., Stein, T. V., Confer, J. J., Gape, L., & Sweeting, M. (2009).Stakeholder mapping for recreation planning of a Bahamian National Park.Society & Natural Resources, 22(2), 111e127.

Fox, E., Poncelet, E., Connor, D., Vasques, J., Ugoretz, J., McCreary, S., et al. (2013).Adapting stakeholder processes to region-specific challenges in marine pro-tected area network planning. Ocean & Coastal Management, 74, 24e33.

Government of Western Australia. (2011). Kimberley science and conservationstrategy. Perth: Government of Western Australia.

Grimble, R., & Chan, M. K. (1995). Stakeholder analysis for natural resource man-agement in developing countries. In Natural resources forum (Vol. 19. No. 2, pp.113e124). Blackwell Publishing Ltd.

Grimble, R., & Wellard, K. (1997). Stakeholder methodologies in natural resource

G. Brown et al. / Applied Geography 67 (2016) 77e93 93

management: a review of principles, contexts, experiences and opportunities.Agricultural Systems, 55(2), 173e193.

Gruby, R. L., Gray, N. J., Campbell, L. M., & Acton, L. (2015). Towards a social scienceresearch agenda for large marine protected areas. Conservation Letters, 00(0),1e11.

Halpern, B. S., Walbridge, S., Selkoe, K. A., Kappel, C. V., Micheli, F., D'Agrosa, C., et al.(2008). A global map of human impact on marine ecosystems. Science, 319,948e952.

Jones, P. J. (2002). Marine protected area strategies: issues, divergences and thesearch for middle ground. Reviews in Fish Biology and Fisheries, 11(3), 197e216.

Juffe-Bignoli, D., Burgess, N. D., Bingham, H., Belle, E. M. S., de Lima, M. G., et al.(2014). Protected planet report 2014. Cambridge, UK: UNEP-WCMC.

Klein, C. J., Chan, A., Kircher, L., Cundiff, A. J., Gardner, N., Hrovat, Y., et al. (2008).Striking a balance between biodiversity conservation and socioeconomicviability in the design of marine protected areas. Conservation Biology, 22(3),691e700.

Lubchenco, J., Palumbi, S. R., Gaines, S. D., & Andelman, S. (2003). Plugging a hole inthe ocean: the emerging science of marine reserves. Ecological Applications, 13,3e7.

Lundquist, C. J., & Granek, E. F. (2005). Strategies for successful marine conservation:integrating socioeconomic, political, and scientific factors. Conservation Biology,19(6), 1771e1778.

Mitchell, R. K., Agle, B. R., & Wood, D. J. (1997). Toward a theory of stakeholderidentification and salience: defining the principle of who and what reallycounts. Academy of Management Review, 22(4), 853e886.

Pollnac, R., Christie, P., Cinner, J. E., Dalton, T., Daw, T. M., Forrester, G. E., et al. (2010).Marine reserves as linked social-ecological systems. Proceedings of the NationalAcademy of Sciences, 107, 18262e18265.

Pollnac, R. B., Crawford, B. R., & Gorospe, M. L. (2001). Discovering factors that in-fluence the success of community-based marine protected areas in the Visayas,Philippines. Ocean & Coastal Management, 44(11), 683e710.

Pomeroy, R., & Douvere, F. (2008). The engagement of stakeholders in the marinespatial planning process. Marine Policy, 32(5), 816e822.

Ramirez, R. (1999). Stakeholder analysis and conflict management. In D. Buckles(Ed.), Cultivating peace: Conflict and collaboration in natural resource

management. Ottawa, Canada: International Development Research Centre.Raymond, C., & Brown, G. (2006). A method for assessing protected area allocations

using a typology of landscape values. Journal of Environmental Planning andManagement, 49(6), 797e812.

Reed, M. S., Graves, A., Dandy, N., Posthumus, H., Hubacek, K., Morris, J., et al. (2009).Who's in and why? A typology of stakeholder analysis methods for naturalresource management. Journal of Environmental Management, 90(5),1933e1949.

Ruiz-Frau, A., Edwards-Jones, G., & Kaiser, M. J. (2011). Mapping stakeholder valuesfor coastal zone management. Marine Ecology Progress Series, 434, 239e249.

Schlossberg, M., & Shuford, E. (2005). Delineating ‘public’ and ‘participation’ inPPGIS. URISA Journal, 16(1), 15e26.

Shucksmith, R., Gray, L., Kelly, C., & Tweddle, J. F. (2014). Regional marine spatialplanningethe data collection and mapping process. Marine Policy, 50, 1e9.

Strickland-Munro, J., Kobryn, H., Moore, S., & Brown, G. (2015a). Human values andaspirations for coastal waters of the Kimberley: Social values and managementpreferences using public participation GIS. Technical Report. Perth, WesternAustralia: Kimberley Marine Research Program Node of the Western AustralianMarine Science Institution.

Strickland-Munro, J., Moore, S., Kobryn, H., & Palmer, D. (2015b). Values and aspi-rations for coastal waters of the Kimberley: Social values and participatory map-ping using interviews. Technical Report. Perth, Western Australia: KimberleyMarine Research Program Node of the Western Australian Marine ScienceInstitution.

Susskind, L., & Thomas-Larmer, J. (1999). Conducting a conflict assessment. InL. Susskind, S. McKearnan, & J. Thomas-Larmer (Eds.), The consensus buildinghandbook (pp. 99e136). Thousand Oaks, CA: Sage.

Thomson-Dans, C., Overman, J., & Moncrieff, D. (2011). Protecting the Kimberleywilderness. Landscope, 27, 32e39.

Voyer, M., Gladstone, W., & Goodall, H. (2012). Methods of social assessment inmarine protected area planning: Is public participation enough? Marine Policy,36(2), 432e439.

Weible, C. M. (2007). An advocacy coalition framework approach to stakeholderanalysis: understanding the political context of California marine protectedarea policy. Journal of Public Administration Research and Theory, 17(1), 95e117.